EP0887511B1 - Method and system for the detection of the precession of a drill string element - Google Patents

Description

The present invention relates to the field of drilling techniques for oil and gas wells, and more specifically detection techniques of the precession during drilling of the element of the drill string commonly called "BHA" for "Bottom Hole Assembly". For simplify, the part of the lining concerned by the present invention will be referred to as BHA in the following description.

BHA is subject to precession when it undergoes a movement proper rotation due in particular to bending. This clean movement is influenced by drilling parameters (weight on the tool, rotation speed, etc.) and geometric parameters (dimensions of the BHA constituents, tool diameter, hole path, type of terrain, etc.). He is considered as a serious malfunction because it prevents the smooth running of the drilling and can even cause significant material damage, particularly to elements of the BHA.

We know in the profession the so-called "whirling" dysfunction (cf. for example the publication "Surface Detection of Vibrations and Drilling Optimization: Field Experience ", Hennense), but it has so far only been taken into account at the level of the drilling tool to reduce breakage and excessive wear on the cutting elements due to to this malfunction.

• acquisition d'au moins une mesure par des moyens de mesure placés proches de la surface, ladite mesure étant représentative de la vibration dudit élément dans ledit puits,
• traitement automatique de ladite mesure comprenant le repérage d'un saut de moyenne et le repérage d'une raie significative dans le spectre de ladite mesure, lesdits repérages apparaissant dans un intervalle de temps inférieur à environ 10 secondes,
• détermination d'un critère de précession à partir des caractéristiques de ladite raie,
• déclenchement d'une alarme si le critère atteint une valeur seuil déterminée.

Thus, the present invention relates to a method of detecting the precession of an element of a drilling rig rotating in a well. The method includes the following steps:

• acquisition of at least one measurement by measurement means placed close to the surface, said measurement being representative of the vibration of said element in said well,
• automatic processing of said measurement comprising locating an average jump and locating a significant line in the spectrum of said measurement, said locations appearing in a time interval of less than approximately 10 seconds,
• determination of a precession criterion from the characteristics of said line,
• triggering of an alarm if the criterion reaches a determined threshold value.

The measurements acquired at the surface can be a function of the couple S (t) and of the voltage V (t) at the top of the drill rods.

The average jump can be identified from the torque measurement S (t) and the location of said significant line from the measurement of the voltage V (t).

The significant line can be included in the frequency interval corresponding to the interval I = [ω / 2π; | Ωb / 2π |], ω being the mean speed of the lining and Ω b = -ω R vs ( R b - R vs ) where Rc is the radius of said precession element and Rb is the radius of the well, and said line can have a power at least greater than 65% of the main maximum.

We can calculate C = Ω / Ωb, Ω being calculated from the frequency of said significant line, and the alarm can be triggered when C is greater to about 0.5, and preferably 0.6.

• des moyens d'acquisition d'au moins une mesure, lesdits moyens de mesure étant placés proches de la partie supérieure de ladite garniture, ladite mesure étant représentative de la vibration dudit élément dans ledit puits,
• des moyens de traitement automatique de ladite mesure comprenant des moyens de repérage d'un saut de moyenne et de repérage d'une raie significative dans le spectre de ladite mesure, lesdits repérages apparaissant dans un intervalle de temps inférieur à environ 10 secondes,
• des moyens de détermination d'un critère de précession à partir des caractéristiques de ladite raie,
• des moyens de déclenchement d'une alarme si le critère atteint une valeur seuil déterminée.

The invention also relates to a system for detecting the precession of an element of a drilling rig rotating in a well. The system includes:

• means for acquiring at least one measurement, said measurement means being placed close to the upper part of said lining, said measurement being representative of the vibration of said element in said well,
• means for automatic processing of said measurement comprising means for locating an average jump and for locating a significant line in the spectrum of said measurement, said locations appearing in a time interval of less than approximately 10 seconds,
• means for determining a precession criterion from the characteristics of said line,
• means for triggering an alarm if the criterion reaches a determined threshold value.

The acquisition means may include measurement sensors related to the torque S (t) and to the voltage V (t) at the upper end of the packing.

The torque S (t) can be measured by means of the motorization in lining rotation and tension can be measured from tension of the cable.

Remember that precession can be defined as movement of the drill collar when the contact of the drill collar against the walls of the well generates an orbital movement of one or more sections of the BHA. It is therefore a composition of two rotational movements.

• la vitesse de rotation imposée au train de tiges par l'appareil de surface (rotation autour de l'axe de la garniture),
• la vitesse propre de rotation due à la précession (rotation autour de l'axe du trou foré),
• V la vitesse de glissement au point de contact (dans le cas où il n'y a pas choc),
• b vitesse de précession sans glissement (V=0),
• Rb rayon du trou, et Rc rayon d'une section en précession,

on obtient la relation V=(Rb-Rc).Ω+Rc. ω, donc pour V=0, on a Ωb=- ω.Rc/(Rb-Rc)Noting:

• the speed of rotation imposed on the drill string by the surface device (rotation around the axis of the packing),
• the natural speed of rotation due to the precession (rotation around the axis of the drilled hole),
• V the sliding speed at the point of contact (in the case where there is no shock),
• b precession speed without sliding (V = 0),
• Rb radius of the hole, and Rc radius of a section in precession,

we obtain the relation V = (Rb-Rc) .Ω + Rc. ω, so for V = 0, we have Ωb = - ω.Rc / (Rb-Rc)

We define s = Ω / Ωb

When Ω is in the same direction as ω, we speak of precession “Forward”, otherwise we speak of “backward” or “retrograde” precession. The synchronous precession (Ω = ω) is the most unfavorable from a point of view abrasion of the drill string, because it is always done in the same place, and that the sliding speed is maximum. On the other hand, in retrograde precession, it there is no specific abrasion phenomenon, but the frequency is high, fatigue is important. In fact, we are in an alternate bending situation which is all the greater as we get closer to Ωb. In practice, the major part of the ruptures and breakdowns are linked to fatigue, and it should therefore be avoided priority the strong retrograde precession, i.e. the case where s = Ω / Ωb approaches 1.

The purpose of detection is to determine Q, to deduce in which precession conditions we find ourselves by studying the value of s, and of trigger an alarm if necessary. Several alarm levels can be defined according to the value of s.

The factors influencing precession are essentially the operational drilling parameters (rotation speed, tool weight, type of mud ...), the geometry of the whole (dimension of the well, inclination of the well, ...) and the type of rock drilled.

Indeed, it appears that the tendency to precession increases when the weight on the tool decreases and the speed of rotation increases. A strong friction between wall and BHA (due to a rough wall, and / or mud poorly lubricating) promotes retrograde precession. We see the same facts in the case of a hard rock, on the one hand, the damping of the interaction tool / rock and vibration is low, on the other hand, the feed speed is generally unimportant.

In addition, the flexural deformation is naturally limited by the diameter drilled.

In deviated wells, it is reduced by damping vibrations at the stabilizers.

Precession is therefore particularly present in wells or parts of vertical or slightly deviated wells.

In practice, most of the breaks and failures are related to the fatigue, and therefore the retrograde precession of the BHA. The remedy is almost always to stop the rotation of the rod to remove the precession. Detecting precession is therefore essential to stop the movement before it has grown too large.

It is also very advantageous to determine the phenomenon of precession from characteristic signals that can be arranged in placing sensors on the upper end of the drill rods linked to the upper part of the BHA. Indeed, the bottom sensors (located at the level of BHA) require heavy means of transmission, since it is essential here to have a large volume of information.

• La figure 1 montre schématiquement une ensemble de forage comportant un système de détection de la précession selon l'invention,
• La figure 2 est un exemple d'organigramme de la méthode,
• La figure 3 est un autre exemple d'organigramme.
• Les figures 4a et 4b montrent la mesure et le spectre de la flexion de la BHA,
• Les figures 5a et 5b montrent la mesure du couple en surface et le spectre de la tension en surface.

The present invention will be better understood and its advantages will appear more clearly on reading the description of an exemplary embodiment, in no way limitative, illustrated by the attached figures, among which:

• FIG. 1 schematically shows a drilling assembly comprising a precession detection system according to the invention,
• FIG. 2 is an example of a flow diagram of the method,
• Figure 3 is another example of a flowchart.
• FIGS. 4a and 4b show the measurement and the spectrum of the bending of the BHA,
• Figures 5a and 5b show the measurement of the surface torque and the surface tension spectrum.

According to the invention, the simplified flowchart of the precession alarm of the BHA is as follows:

• des moyens d'acquisition desdites mesures qui font transiter les signaux correspondant vers :

• des moyens de traitement automatiques des données qui déterminent le déclenchement ou non :

• d'une alarme.

A set of sensors provides characteristic measurements of the lateral vibrations of the BHA, (e.g. torque and tension) at:

• means for acquiring said measurements which pass the corresponding signals to:

• automatic data processing means which determine whether or not it is triggered:

• an alarm.

The main objective of the invention is to detect precession from only signals available on the surface. Now, the bending phenomena are not very sensitive on the surface because on the one hand, there is such a dissipation bending waves (structural and due to the well fluid) that the latter tend to to disappear before reaching the surface, and on the other hand, the well is a bad guide for bending waves.

However, there are many couplings between the different modes (longitudinal, torsion, bending). So we can find signs precession characteristics in surface signals other than the bending.

Figure 1 depicts the drilling of a well 1 using a surface 2 comprising a tower 3 and suspension means 4 to which is attached to the drill string 5. The drill string 5 is composed as it is known in the art of drilling, by a drilling tool 6, a set of rods 7 in which one or more stabilizers 8 can be integrated, drill rods 9 screwed together to make the connection mechanical with the injection head, for example motorized 10, suspended from lifting means 4. It is the precession of part of BHA 7 which concerns the present invention.

A connector 11 includes sensors sensitive to the vibration of the train rods. The vibrations recorded can be torsion, bending, tension, or even pressure in the drilling fluid injected through the head 10. In this example, the sensors are sensitive to the torque and at the tension at the upper end of the rods 9. For the measurement torque, suitable sensors can be installed in the setting-up means gasket rotation: rotary table or motorized injection head 10. The tension can also be measured from the tension of a strand of the cable drilling.

Means of acquisition and processing 12 of the corresponding signals to the vibration of the torque and the surface tension, respectively (S (t) and V (t), are in connection with the instrumented fitting 11, either by a cable 13 or by radio wave received by the antenna 14. A drilling control cabin 15 receives, either by radio means 16 or by cable 17, the information from the acquisition and processing means 12. This information can trigger signaling means 18 of the alarm type, by example of lights of different colors depending on a specific level of precession.

S(t) = couple en surface,

V(t) = tension en surface.

The automatic processing of data and the alarm uses in the present example as signals characteristic of lateral vibrations:

S (t) = surface torque,

V (t) = surface tension.

The automatic data processing is done according to the following scheme:

ω représente la vitesse de rotation instantanée imprimée à l'ensemble de forage par le moteur.

We recall that:

ω represents the instantaneous rotation speed imparted to the drilling assembly by the engine.

Ωb is the precession rate without sliding. It is calculated from the following expression: Ω b = -ω R vs ( R b - R vs ) where Rc is the radius of the BHA and Rb is the radius of the well.
t _am is the time when an average jump of the signal studied in time is detected, here the surface torque S (t).

• « Détection of Abrupt Changes. Theory and Application » par Basseville et Nikiforov-Prentice Hall Information and System Sciences Series, 1993.
• « Digital Processing of Random Signals. Theory and Methods » par Porat- Prentice Hall Information and System Sciences Series, 1994.

The following documents are cited here with reference:

• “Detection of Abrupt Changes. Theory and Application ”by Basseville and Nikiforov-Prentice Hall Information and System Sciences Series, 1993.
• "Digital Processing of Random Signals. Theory and Methods ”by Porat- Prentice Hall Information and System Sciences Series, 1994.

Detection of the jump from average to temporal:

An example of automatic processing for detecting the jump of average of the signal processed in time (for example here, the couple).

To detect the jump in average torque, we can choose to use sliding windows, one large (for example 2000 points) and the other of small dimension (for example 400 points) which contains the points newer ones from the big window. We then compare the average obtained in the large window with that obtained by the small window. At the time of change in mean, the two results become appreciably different, because the small window is very sensitive to change while the large is much less so. This comparison can be made from log likelihood ratio of the two means, calculated from continuously using the CUSUM algorithm. The CUSUM algorithm was described by M. Page in 1954 in "Continuous Inspection Schemes" Biometrika, Vol. 41, PP 100-115.

The CUSUM algorithm makes it possible to give precisely the time t _am when there is a change in the mean of the signal.

Spectral change detection:

From the moment when we know the instant t _am of the change in time mean of the torque, we can calculate the spectrum of the voltage signal before this instant and the spectrum after this instant and compare them. These spectra are useful on the frequency band I = [ω / 2π; | Ωb / 2π |] mentioned in the preceding flowchart. It is in this interval that we seek to detect the appearance of a line.

1) Utiliser à nouveau la comparaison, en fréquentiel cette fois, entre une petite fenêtre et une grande fenêtre comme définies précédemment. Pour cela, il faut connaítre suffisamment précisément les caractéristiques fréquentielles du signal. On modélise donc au préalable le signal par une modélisation autorégressive. Ensuite, on peut calculer la distance spectrale entre les deux spectres des deux fenêtres, en utilisant à nouveau l'algorithme CUSUM.

2) Faire une comparaison directe des spectres avant et après t_am. Cette méthode consiste à faire la différence directe entre les 2 spectres ainsi calculés, en admettant une petite fluctuation de la valeur de la fréquence associée à chaque raie.

To compare them, two solutions are proposed here:

1) Use the comparison again, in frequency this time, between a small window and a large window as defined above. For this, it is necessary to know sufficiently precisely the frequency characteristics of the signal. One thus models the signal beforehand by an autoregressive modeling. Then, we can calculate the spectral distance between the two spectra of the two windows, using the CUSUM algorithm again.

2) Make a direct comparison of the spectra before and after t _am . This method consists in making the direct difference between the 2 spectra thus calculated, by admitting a small fluctuation of the value of the frequency associated with each line.

• Si la réponse est non, il n'y a pas déclenchement d'alarme et on se borne à continuer à repérer un éventuel changement de moyenne temporelle dans le signal de couple.
• Si la réponse est oui, on détermine alors la puissance de cette raie par rapport aux autres, afin de déterminer si elle est significative. Un pourcentage spécifique de la puissance totale du signal est défini comme palier pour sélectionner les raies « significatives ».

Whatever solution is chosen, the comparison of the two spectra allows us to answer the following question:
"At time t _am , was there a line in the interval I?"

• If the answer is no, there is no alarm triggering and we limit ourselves to continuing to identify a possible change in time average in the torque signal.
• If the answer is yes, we then determine the power of this line compared to the others, in order to determine if it is significant. A specific percentage of the total signal power is defined as a stage for selecting the "significant" lines.

So, at the end of automatic data processing, we selected lines considered important and which may be evidence of the presence of a precession movement of the BHA. The frequency of the line significant allows to calculate the precession speed Ω.

The alarm can be described according to the following diagram:

A criterion, denoted C, for characterizing the precession is defined as follows:
C = Ω / Ωb, where Ω is the proper rotation speed of the BHA in its precession movement.

We also define a threshold s such that if C> s, then the precession considered is a retrograde precession, and therefore very harmful to the lining drilling.

An alarm is therefore triggered insofar as the calculation of criterion C and its comparison with the threshold s allows us to conclude on the appearance of a BHA retrograde precession movement.

Figures 2 and 3 illustrate two more detailed algorithms of the automatic data processing and alarm triggering based on the principles of precession detection described above.

According to the flow diagram of FIG. 2, spectral processing is only carried out from the moment when an instant t _am has been determined when an average jump has been detected (block 30). Block 40 represents the frequency treatments on upstream and downstream windows with respect to t _am .

In this block 40, an estimate of the spectrum of the front torque is calculated and after the mean jump by conventional signal processing methods (for example: averaged periodograms, high resolution methods, ...). We also calculates the average of the speed of rotation considered as stationary when whirling appears. We then assess the difference of the spectra thus obtained and one keeps, for example, the four most energetic lines.

In parallel (case of Figure 2) or sequentially (case of Figure 3), we calculates Ωb from ω and the physical characteristics Rb and Rc.

Finally, we select one or more lines by sorting the different spectral lines in energy, for example at 65% compared to maximum energy, and in frequency by keeping only the frequencies belonging to the interval. I = [ω / 2π; | Ωb / 2π |].

Following block 40, C is calculated for the significant line.

In this example, the alarm threshold is taken at 0.6. If C is greater than 0.6, an alarm is triggered to prevent the significant risk of having a dysfunction close to the retrograde type.

FIG. 3 shows another example of automatic processing in which the monitoring of the mean and the spectral content is done in parallel over time (block 30 ′). In the case where, there is jump of average at time t _am and spectral change at time t _as and these two instants are sufficiently close, for example if | t _am -t _as | is less than 8 seconds, we then trigger the search for a significant line (block 40 '). We then find the same functionality for triggering an alarm.

These different procedures are only examples of triggering automatic alarm regarding the precession of the BHA but others procedures are possible, including, for example, automatic processing based on Bayesian statistics.

Figures 4a, 4b, 5a and 5b are comparative examples of what is can actually measure at the BHA level and what we measure in area. FIG. 4a is the recording of the flexion recorded at the level drill collars, for example in fitting 20 (Figure 1). Such means of measurement are for example described in document EP-B1-0558379, cited here in reference. We note the jump in average referenced 21.

FIG. 4b gives the spectrum corresponding to the signal of the moment of bending at the BHA level. The continuous curve 23 corresponds to the spectrum in upstream of the average jump, i.e. without whirling, the dotted curve 24 corresponds to the spectrum after the average jump, that is to say in whirling. We note the significant referenced line 25 because in the frequency range 15-20 Hertz and characteristic power.

At the same time, the vibrational signals were recorded on the surface torque (Figure 5a) and the spectrum of surface tension (Figure 5b). We well find the jump from average 31 actually corresponding to the BHA whirling and significant line of the same frequency as the frequency whirling directly recorded at the bottom.

Claims (8)

1. A method for detecting precession of an element of a drill string (7) in rotation in a well (1), comprising

   acquisition of at least one measurement by measuring means placed close to the surface, said measurement being representative of the vibration of said element in said well,

   the method being characterized in that it comprises the following stages :

   automatic processing of said measurement, comprising locating a jump in average and locating a significant line in the spectrum of said measurement, said locations appearing in a time interval shorter than about 10 seconds,

   determination of a precession criterion from the characteristics of said line,

   setting off of an alarm if the criterion reaches a determined threshold value.
2. A method as claimed in claim 1, wherein said measurements acquired at the surface depend on the torque S(t) and on the tension V(t) at the top of drillpipes (5).
3. A method as claimed in claim 2, wherein the jump in average is detected from the measurement of torque S(t) and wherein said significant line is determined from the measurement of tension V(t).
4. A method as claimed in any one of the previous claims, wherein said significant line is in a frequency interval corresponding to interval I = [ω/2π,|Ωb/2π|], ω being the average speed of the string and Ω b = -ω Rc (Rb - Rc ) , where R_c is the radius of said element under precession and R_b is the radius of the well, and said line has a power at least greater than 65 % of the main maximum.
5. A method as claimed in any one of the previous claims, wherein C = Ω/Ω_b is calculated, Ω being calculated from the frequency of said significant line, and wherein said alarm is set off when C is greater than about 0.5, preferably greater than 0.6.
6. A system for detecting precession of an element of a drill string (7) in rotation in a well (1), comprising

   means (11) for acquisition of at least one measurement, said measuring means being placed close to the upper part of said string, said measurement being representative of the vibration of said element in said well,

   the system being characterized in that it comprises :

   means (12) for automatic processing of said measurement, comprising means for locating a jump in average and for locating a significant line in the spectrum of said measurement, said locations appearing in a time interval shorter than about 10 seconds,

   means for determining a precession criterion from the characteristics of said line,

   means (18) for setting off an alarm if the criterion reaches a determined threshold value.
7. A system as claimed in claim 6, wherein the acquisition means comprise measurement detectors for measurements linked with the torque S(t) and the tension V(t) at the upper end of string (5).
8. A system as claimed in claim 7, wherein torque S(t) is measured by the rotating motive means of string (5) and wherein the tension is measured from the tension of the cable.

Images